Archi/e Machina: Interactive ArchitectureBased on Tensegrity

Yosuke UshigomeTokyo, Japan

ushi@cyber.t.u-tokyo.ac.jp

Ryuma NiiyamaGraduate School of Information

Science and Technology

The University of Tokyo

niiyama@isi.imi.i.u-tokyo.ac.jp

Kunihiro NishimuraGraduate School of Information

Science and Technology

The University of Tokyo

kuni@cyber.t.u-tokyo.ac.jp

Tomohiro TanikawaGraduate School of Information

Science and Technology

The University of Tokyo

tani@cyber.t.u-tokyo.ac.jp

Michitaka HiroseGraduate School of Information

Science and Technology

The University of Tokyo

hirose@cyber.t.u-tokyo.ac.jp

Abstract—This paper elucidates the possibilities of in-teractive architectural-space. Interactive architectural-space isarchitectural-space that is changeable or adaptable in interactionwith internal and external users and environments. This paperilluminates the point which has not been focused on and describesa methodology of making a prototype of interactive architectural-space to investigate the point. Designed prototype is experimentedand evaluated with psychological methods.

I. INTRODUCTION

In this section, background and purpose of this research is

described.

A. Interactivity in Contemporary Architecture

Human-beings do various activities involved with

architectural-space and architects have designed various

architectural-space through history. Modernism architecture is

one of the biggest paradigms in architectural design and it

seem to still carry on. However, this style of design, built with

concrete and glass, is often criticized as design without human

embodiment, transformation or development. Our lifestyles,

natural environment are changing so drastically that these

modernism architectures no longer have the adaptability to

them.

Many contemporary architect attempt to solve this prob-

lem and create a new paradigm of architecture; Ito’s ambi-

ent, ephemeral architecture and Kuma’s organic-material-based

architecture are both at the forefront of it (Fig. 1). These

contemporary architectures set ambiguous boundaries between

human behaviors and surrounding environments, leaving room

for spatial and temporal transformation. In this sense, contem-

porary architecture can be said as the interactive visualization

of the relationships between human and environment.

B. Interactive Architecture by Cybernetics

Cybernetics have theoretically researched on its applications

into architectural-space since 1960’s. Pask regarded an architec-

tural environment as an adaptive system that developed through

controlling users and being controlled by users and illuminated

a superordinate interaction in a design process of such system,

where the system and its designer control each other[1]. He

also presented an idea that an environment analyzes and adapts

to patterns of inhabitants’ behavior using computer.

Recent advancement in information technology has increased

the feasibility of these ideas that could not be actualized

because of the limitation in technology and in economic issues.

These ideas and theories on cybernetics will contribute to

already-described concept of contemporary architecture.

C. Purpose of Research

On this background, many researches on interactive

architectural-space that is changeable and adaptable are per-

formed nowadays[2][3]. They deal with various architectural-

space in its form, scale, method but all of them attempt to

seek and suggest future styles of human beings, architectures,

environments.

When we see interactive architectural-space not as reactive

mechanical system but as a new frontier of the history of

architectural-space, its spatial nature matters; in especially

architectural-space that physically moves, the spatial perception

of inhabitants will change drastically. Despite that, there is

no researches on changes in spatial nature or perception and

there is no prototypes that can change and transform whole

architectural-space. Purposes of this research is to make a

prototype of interactive architectural-space that can move to

change its spatial nature and to seek influences by its interaction

on spatial experience and perception.

D. Composition of This Paper

In section I, the architectural and cybernetic background and

the purposes of this research were described. In section II,

relevant researches is introduced to position this paper clearly.

In section III, a method adopted to design a prototype of

interactive architectural-space is introduced. In section IV, a

978-1-4244-9026-4/10/$26.00 ©2010 IEEE - 55 -

(a) Ito’s Sendai Mediatheque (b) Kuma’s Great (Bamboo) Wall

Fig. 1. Contemporary Architectures

designed prototype and its specification are described. In sec-

tion V, experiments on prototype’s spatial nature are performed

and its result is shown. In section VI, a conclusion and horizons

of this research are described.

II. RELEVANT RESEARCH

Section II summarizes existing researches on interactive

architectural-space, illuminates the point that has not been

discussed by them and positions this research.

A. Interactive Architectural-space

In this paper, interactive architectural-space means

architectural-space that is changeable or adaptable in

interaction with internal and external users and environments.

The methods of interaction are implemented in many different

ways and this subsection introduces each methods; interaction

by audio and visual equipment, interaction by robotics and

interaction by spatial movement.

1) Interaction by Audio and Visual Equipment: Interaction

by spatial visual equipments has been developed mainly by

virtual reality technology and related researches[4]. Recently,

based on the widespread of these technologies, many attempts

to expand interaction by audio and visual equipments into real

architectural-space are performed. The building of Ars Elec-

tronica Center, the world’s biggest center of art and technology,

has 100m2 wall whose whole surface is embedded full-colored

LEDs and can transform their color according to real-time

control(Fig. 2).

However, these interactions by audio and visual equipments

seem to lack spatial transformability.

2) Interaction by Robotics: According to the definition of

robot as machine in which controllers integrated sensors and

actuators, there have been many robotics in architectural-

space; automatic doors, lifts, automatic lighting equipments

etc. Moreover, recent development in robotics enabled more

technological robots to enter into our living space, such as

cleaning robots, robot pets, surgical robot etc. Using these

robotics, Sato and colleagues suggested a concept of “Robotic

Room” and its implementation(FIg. 3). They also built up

Fig. 2. “LIGHTS ON!” performance by YesYesNo

Fig. 3. “Real-World-Showroom” by the Univ. of Tokyo

a methodology on interactions between human and machine

behaviors in a room equipped with manipulators[5].

But this approach is not to interact with architectural-space

itself but to dispose interactive robots or manipulators in

existing architectural-space. In this sense, it has a same defect

as audio and visual approach.

3) Interaction by Spatial Movement: Robotics is getting so

stable as stated and so detail structural analysis is built up that,

recently, even architectural-space itself can move and change

its spatial nature. It is essential in interactive architectural-space

that architectural structures and spatial boundaries interact with

- 56 -

(a) Adaption by movement (Nouvel, xxxx) (b) Expression by movement (Beesley, 20xx)

Fig. 4. Various movement of interactive architecture

behaviors of users and environments. Fig. 4 shows examples

for spatial movement in architectural-space. Nouvel designed

“l’Institut du monde arabe” shown in Fig. 4(a). Its wall has

geometric patterns that mechanically moves to adjust sunlight

from outside into architectural-space. Its gimmicks not only

alter lighting condition of the space but also affect on in-

habitants’ perception of spatial extensity[6]. Fig. 4(b) shows

“Hylozoic Soil” by Beesley. This architecture-scale installation

reacts human inside and moves a part of post near human. It

expresses biotic interaction in its very slow, like sea animal,

movement.

This approach has an amount of potential in changing

user’s spatial perception. Especially, when architectural-space

involving human moves in synchronism with movement of

inhabitants, it is to be expected that the spatial perception of

inhabitants alter drastically. Despite that, existing prototypes or

artworks can move only a part of whole space that isn’t enough

to investigate human perception precisely. A prototype of

interactive architectural-space that can interact with inhabitants

by moving its whole space has to be created and investigations

of its influence on human perception has to be performed.

B. Positioning of This Research

As stated in this section, there are many methods for realizing

interactive architectural-space and interaction by spatial move-

ment has a particular potential in influencing spatial perception.

On the other hand, the evaluation of spatial perception in

existing architectural-space has been generally performed with

Semantic Differential Method (SDM) etc. Jo and colleagues

verified the comfortability of personal space with SDM[7].

This kind of researches on psychological influence and design

method of architectural-space is called environmental psychol-

ogy. Maki and colleagues elucidated the entire picture of envi-

ronmental psychology by describing environmental affordance

and becoming personal space of environment[8]. However,

these methodology in environmental psychology have yet to

be applied to computerized architectural-space.

In this paper, a new prototype for interactive architectural-

space that can interact with inhabitants by moving its whole

space is introduced and, to evaluate the possibilities of in-

teraction by spatial movement, the methods of environmental

psychology are applied.

Fig. 5. 6 struts tensegrity

III. DESIGNING INTERACTIVE STRUCTURE

In this section, a methodology of actualize an interactive

architectural-space that can change spatial nature and spatial

perception by movement. Firstly, tensegrity structure is intro-

duced, that is an adopted structure because of its movability

in whole space and its long-term and short-term adaptability.

Then, the design of tensegrity structure and its simulation are

described.

A. Tensegrity Structure

To evaluate effects of an interactive architectural-space, it is

essential that the prototype can move not in a part of the space

but in whole spatial structure. This research adopts tensegrity

structure for the spatial structure of the prototype of interactive

architectural-space. Tensegrity has movability in whole part and

long-term and short-term adaptability.

Tensegrity is a structure that was invented 1950’s. The

definition of tensegrity by A. Pugh is as follows: “Tensegrity

systems are spatial reticulate systems in a state of self-stress. All

their elements have a straight middle fibre and are of equivalent

size. Tensioned elements have no rigidity in compression and

constitute a continuous set. Compressed elements constitute a

discontinuous set. Each node receives one and only one com-

pressed element”[9]. Fig. 5 shows the most famous tensegrity

that is strictly based on the definition and has 6 struts and 24

cables.

Because tensegrity is in an equilibrium state between ten-

sional force and compressional force, it can provide a light,

stiff and large space with struts rigid enough to endure compres-

sional force. It has also a spherical network of tensional force

- 57 -

(a) Equilibrated state (b) Unequilibrated state

Fig. 6. Change in equilibrium state

Fig. 7. Designed 21 struts tensegrity dome

TABLE ISTRUCTURAL PARAMETERS OF DESIGNED TENSEGRITY

Parameters Values

Number of Struts 21

Number of Cables 84

Number of Nodes 42

Cables per Nodes 4

Strut’s Length per Cable’s Length 1.852

so that a geometrical change in a small part of the structure,

often by change in the length of the component, spread into

whole part of the structure(Fig. 6). In this way, tensegrity is

suitable for the structure adaptive with human body and with

environments.

B. Designing Interactive Structure

This subsection shows the process of designing and verify-

ing the structure of interactive architectural-space. Tensegrity

structure of the prototype was designed based on principles

bellow:

• Dome shape for spatial extensity enough for user to enter

and move around

• Definition based structure for ease in simulating

• Less components for ease in prototyping and constructing

Fig. 7 shows the model of the designed tensegrity and Table

I its parameters. This structure is tensegrity with 21 struts

and completely based on the definition. Its shape of cut sphere

enables users to enter and move around.

Fig. 8. Diagram of pneumatic muscle

TABLE IIMOTION PARAMETERS OF PNEUMATIC MUSCLE

Parameters Values

Length at normal pressurs 1.35 m

Length at 0.6 MPa 1.08 m

Generative force 100+ N

Mass 50 g

C. Actuation

There are several ways to actuate tensegrity and even some

robots whose structure is tensegrity. Almost all precedents

implement motors on the nodes of tensegrity[10]. However,

the motor-approach makes the whole structure very heavy

when applied into spatial extensity. Instead, our prototype

adopt pneumatic muscle to actuate compressional components

of tensegrity because pneumatic muscle is very light but can

generate strong force under certain pressures[11]. Fig. 8 shows

a diagram of pneumatic muscle and Table II its motion param-

eters.

D. Simulation

To verify the movement of tensegrity with pneumatic mus-

cles, dynamical simulation was performed using Open Dy-

namics Engine. Pneumatic muscles are implemented instead of

corresponding tensional components. The places of pneumatic

muscles are determined according to the tensional force on

corresponding tensional components. The more force acts on

a component, The more the substitute pneumatic muscle elon-

gates. So tensional components are replaced with pneumatic

muscles from the strongest-force-acted one to the weakest-

force-acted one as shown in Fig. 9.

The number of pneumatic muscles are decided according to

the safe height of collapsing state in which all pneumatic mus-

- 58 -

Fig. 9. Distribution of tensional force

Fig. 10. Change in height of the structure

TABLE IIISIMULATION PARAMETERS

Parameters Values

Length of strut 4000 mm

Mass of strut 0.2 kg

Natural length of cable 2140 mm

Spring constant of cable 300 N/m

Mass of cable 0 kg

cles are elongated. 12 pneumatic muscles make the minimum

height 2.32 m which is safe for an ordinary adult not to touch

or hang on components as shown in Fig. 10.

Using these conditions dynamical simulation was performed.

Fig. 11 shows the screen captures of dynamical simulation.

Fig. 11(a) represents a state where all pneumatic muscles are

elongated and Fig. 11(b) contracted. In Fig. 11, colored lines

represent pneumatic muscles; blue ones are elongated and red

ones are contracted. Table III shows simulation parameters.

IV. ARCHI/E MACHINA

In this section, the detail information of the prototype of

interactive architectural-space is described. The design of the

prototype is based on the method stated in section III. This

prototype is named “Archi/e Machina”. Table IV shows all

TABLE IVARCHI/E MACHINA’S PARAMETERS

Structural Parameters Values

Number of Struts 21

Number of Cables 84

Number of Nodes 42

Cables per Nodes 4

Strut’s Length per Cable’s Length 1.852

Material of Strut Aluminum

Length of Strut 4000 mm

Mass of Strut 0.4 kg

Material of Cable Steel Wire

Length of Cable 2140 mm

Mass of Cable 0.1 kg

Total Weight 18 kg

Moving Parameters Values

Generative Force Needed 25 N

Number of Pneumatic Muscle 12

Natural Length of Pneumatic Muscle 2410 mm

Mass of Pneumatic Muscle 0.05 kg

Maximum Height of Structure 3460 mm

Minimum Height of Structure 2320 mm

Fig. 12. System overview of Archi/e Machina

parameters concerned with designing and constructing Archi/e

Machina.

Fig. 12 shows the system overview of Archi/e Machina.

Environmental information such as human behavior and envi-

ronmental parameters are sensed and analyzed by appropriate

sensors and PCs. Then each pneumatic muscles is controlled

independently by pneumatic valves using micro-controllers and

electric circuit. The pneumatic valve can control the direction

and the speed of airflow so that the whole space can move very

complexly according to environmental information.

A. Half-scale Prototype

Firstly, a half-scale prototype was made. Fig. 13 shows the

half-scale model of Archi/e Machina. This model was exhibited

at Haneda Airport in Tokyo as one of artworks in “Air Harbor

- Digital Public Art Exhibitoin”. People could not enter the

space because of its scale so that the acoustic information

in the airport or by people was used as the “Environmental

information” in Fig. 12. People enjoyed seeing that Archi/e

Machina reacted to the announcement of air port and interacting

- 59 -

(a) Collapsing State (No muscles contract) (b) Rising State (12 muscles contract)

Fig. 11. Simulation of dynamic movement

Fig. 13. Archi/e Machina (Half Scale)

Fig. 14. Archi/e Machina (Full Scale)

with clapping, jumping and speaking.

B. Full-scale Prototype

Based on the succeed of making half-scale model, full-scale

model of Archi/e Machina was actualized as shown in Fig. 14.

Full-scale model has the same parameters as described in Table

IV and uses Infra-red sensors to detect human movement inside

itself. People can enter and move around the space that react

to their behavior. Fig. 15 shows the example of the movement

of Archi/e Machina.

As shown in this section, “Archi/e Machina”, a new prototype

for interactive architectural-space that can interact with inhab-

itants by moving its whole space, was successfully realized.

Fig. 16. Scheme of experiment

V. EXPERIMENTS AND RESULTS

In this section, the prototype “Archi/e Machina” is evaluatee

with the methods of environmental psychology.

A. Psychological Experiment

Semantic Differential Method (SDM), which was generally

used in environmental psychology was applied. SDM evaluates

stimuli with multiple adjective couples to illuminate psycho-

logical images. In SDM, factor analysis is used for analyzing

data described by subjects.

Fig. 16 shows the scheme of the experiment and adjective

couples. There are 4 patterns in interactive spatial experiences

as is explained FIg. 17. This experiment was performed with

13 subjects.

B. Results and Discussion

Fig. 18 shows the results of the image evaluation experiment,

representing the changes of factors’ score of evaluation depend

on interaction patterns. Principal factor analysis gave two

evaluating factors interpreted as below:

• “Rest” for “Boring”, “Passive”, “Stable”, “Old”, “Simple”

and “Calm”

- 60 -

(a) Interacting with human inside (b) Static state

Fig. 15. Full-scale Archi/e Machina

Fig. 17. Interaction pattern in experiment

Fig. 18. Interaction pattern in experiment

• “Tense” for “Constrained”, “Narrow”, “Hard”, “Bad”,

“Fearful” and “Tense”

Fig. 18 indicates that, depend on interaction patterns, the

scores of two factors are radically change and spatial perception

of subjects and spatial nature of Archi/e Machina trasforme.

1) “Rest” Factor: Firstly, all patterns except for pattern 0

resulted in negative “Rest” factor. This result must be because

of the movement of spatial structure itself. Secondly, “Rest”

factor in pattern 2 resulted in quite negative value. It seems

natural because pattern 2 is the only pattern in which Archi/e

Machina moves in reaction with subjects and subjects were

asked to move around to feel spatial nature.

2) “Tense” Factor: “Tense” factor in patter 2 has com-

paratively lower value than in pattern 0 and 1, implying

that synchronized movement of surrounding architectural-space

decreases negative images such as “constrained”, “narrow”

and “fearful”. This synchronization may give subjectives with

active comfortability. Comparing pattern 2 and 3, pattern 3

has more unfavorable “Tense” factor. The reason may be that

the control of surrounding architectural-space is apart from

the subject and is determined by the accidental behavior of

external passengers. This presumption leads to a hypothesis

that surrounding architectural-space that moves interactively

and synchronously can be felt as an augmentation of personal

space so called in psychology[12].

VI. CONCLUSION

A. Summery of This Paper

This paper introduced increasing researches on interactive

architectural-space and elucidated the importance of influence

on spatial perception and spatial nature by interaction be-

tween human and architectural-space. A prototype of interactive

architectural-space was made to investigate the influence. The

prototype adopted tensegrity for the structure of the prototype.

The image evaluation experiment indicated that surrounding

architectural-space that moves interactively and synchronously

can be felt as an augmentation of personal space.

B. Future Work

1) Evaluation Method: The evaluation of spatial perception

in this paper may be inadequate. It has to be scrutinized how

subject’s personal space change according to parameters such

as interaction patterns , speed of actuation, etc. If surrounding

architectural-space was a kind of personal space, it would be

possible to experiment the psychological influence by the conti-

guity to the space by other people and by external environments.

- 61 -

To compare these interactive architectural-space with existing

robotics-based interactive room is also important.

In this paper, experiment used SDM that is generally

used for evaluating existing architectural-space. However,

when architectural-space itself moves, environmental param-

eters change drastically. To evaluate interactive architectural-

space precisely, we have to establish a new evaluation method

that include a point of temporal transformation of architectural-

space.

2) Prototype: The prototype in this paper is just a structure

and is not physically isolated with external environment. It is

obvious that this structural space is not appropriate to real living

environment. A prototype involved with a membrane has to be

made and evaluated with more ordinary human activities.

ACKNOWLEDGMENT

This research is partly supported by JST CREST ”Technol-

ogy to Create Digital Public Art” project.

REFERENCES

[1] Gordon Pask, The Architectural Relevance of Cybernetics, ArchitecturalDesign, September, 1969, pp. 494-496.

[2] John Frazer, An Evolutionary Architecture, Architectural AssociationPublicatoin, 1995.

[3] Usman Haque, Architecture, Interactions, Systems, AU: Arquitetura &Urbanismo 149, 2006.

[4] M. Hirose, T. Ogi, S. Ishiwata and T. Yamada, Development and evaluationof immersive multiscreen display CABIN, Systems and Computers inJapan, Vol. 30, No. 1, pp. 13-22, 1999.

[5] Tomomasa Sato and Taketoshi Mori, Robotic Room: Its concept andrealization, Intelligent autonomous systems: IAS-5, pp. 415-422, 1998.

[6] Selma al-Radi, Institut du Monde Arabe, Architecture for Islamic SocietiesToday, pp. 138-147, 1994.

[7] Hitoshi Joh and Hiroyuki Yoshida, A Study of Amenity in Personal Space,Bulletin of The Faculty of Human Development, Kobe University, Vol. 1,No, 2, pp. 149-163, 1994.

[8] Kiwamu Maki, Environmental Psychology - perspective of environmentalDesign, Shumpu-sha, 2004.

[9] Reno Motro, Tensegrity: Structural Systems for the Future, Butterworth-Heinemann, 2003.

[10] Chandana Paul, Francisco J. Valero-Cuevas, and Hod Lipson, Designand control of tensegrity robots for locomotion, IEEE Transactions onRobotics, Vol. 22, No. 5, pp. 944-957, 2006.

[11] Ryuma Niiyama, Akihiko Nagakubo and Yasuo Kuniyoshi, Mowgli: Abipedal jumping and landing robot with an artificial musculoskeletal sys-tem, 2007 IEEE International Conference on Robotics and Automation,pp. 2546-2551, 2007.

[12] Robert Sommer, Personal Space. The Behavioral Basis of Design., 1969.

- 62 -